feat(broad-3/log-fc): share crunch1 model huggingface link (#31)

competitions/broad-3/quickstarters/log-fc/log-fc.ipynb

@@ -2339,19 +2339,26 @@
     "1.   Prepare the necessary directories and load the configuration parameters from the previously trained model.\n",
     "2.   Provide a dataset as a dict containing at least an H&E image of tissue regions (`HE`) and a nuclear segmentation mask (`HE_nuc`).\n",
     "3.   Preprocess the data into image patches (X_test).\n",
-    "4.   **Load locally `pytorch_model.bin`** to generate embeddings for the dataset\n",
-    "5.   **Load locally the trained regression models** (`config.json` and `.pkl` files from Crunch 1 in `./resources`) for regression predictions on the 460 measured genes.\n",
+    "4.   **Load locally `pytorch_model.bin`** in `./resources` to generate embeddings for the dataset\n",
+    "5.   **Load locally the trained regression models** (`config.json` and `.pkl` files from Crunch 1 in `./resources`) to apply regression predictions on the 460 measured genes.\n",
+    "\n",
+    "`pytorch_model.bin`, `config.json` and baseline regression model from Crunch 1 quickstarter are available on Hugging Face: https://huggingface.co/Tarandros/Autoimmune_Disease_Machine_Learning_Crunch_1/tree/main (Use your own config file and model if you can)\n",
+    "\n",
+    "**Note: You can also manage this Crunch 3 challenge without Crunch 1 & 2: \"For those who did not participate in Crunch 1 and 2, you can design\n",
+    "a gene panel from scratch using biological understanding or other approaches.\"**\n",
     "\n",
     "![train_architecture](https://raw.githubusercontent.com/crunchdao/competitions/refs/heads/master/competitions/broad-1/quickstarters/resnet50-plus-ridge/images/train_architecture.png)\n",
     "\n",
     "**Crunch 2:**\n",
+    "\n",
     "5.   Apply cosine similarities between predictions and single-cell RNA sequencing from shared genes.\n",
     "6.   Compute weighted average of unmeasured genes from scRNA-Seq based on similarity scores.\n",
     "7.   Return the predictions of the 18615 genes (458 shared genes + 18157 unmeasured genes).\n",
     "\n",
     "![similarity_short_example](https://raw.githubusercontent.com/crunchdao/competitions/refs/heads/master/competitions/broad-2/quickstarters/cosine-similarity/images/similarity_short_example.png)\n",
     "\n",
-    "**Note: You need to upload in `./resources` (model_directory_path) the pre-trained Resnet50 `pytorch_model.bin` ([link](https://huggingface.co/timm/resnet50.tv_in1k/tree/main)), the config file `config.json` and your trained regression models (expl: `model_split0.pkl`, `model_split1.pkl`, ...) from Crunch 1 ([link](https://colab.research.google.com/github/crunchdao/quickstarters/blob/master/competitions/broad-1/quickstarters/resnet50-plus-ridge/resnet50-plus-ridge.ipynb)).**"
+    "**Note: You can also manage this Crunch 3 challenge without Crunch 1 & 2: \"For those who did not participate in Crunch 1 and 2, you can design\n",
+    "a gene panel from scratch using biological understanding or other approaches.\"**"
    ]
   },
   {
@@ -2828,7 +2835,7 @@
    "source": [
     "Preprocessed datasets are saved in `/tmp/processed_dataset`(temporary storage - not needed for inference).\n",
     "\n",
-    "**Note: You need to upload in `./resources` (model_directory_path) the pre-trained Resnet50 `pytorch_model.bin` ([link](https://huggingface.co/timm/resnet50.tv_in1k/tree/main)), the config file `config.json` and your trained regression models (expl: `model_split0.pkl`, `model_split1.pkl`, ...) from Crunch 1 ([link](https://colab.research.google.com/github/crunchdao/quickstarters/blob/master/competitions/broad-1/quickstarters/resnet50-plus-ridge/resnet50-plus-ridge.ipynb)).**\n",
+    "**Note: You need to upload in `./resources` (model_directory_path) the pre-trained Resnet50 `pytorch_model.bin`, the config file `config.json` and your trained regression models from Crunch 1 (baseline quickstarter model: https://huggingface.co/Tarandros/Autoimmune_Disease_Machine_Learning_Crunch_1/tree/main (Use your own config file and model if you can)).**\n",
     "\n",
     "### Crunch 1 - Crunch 2 - Crunch 3 -> Gene Ranking"
    ]